Pete Warden is Co-Founder and CEO at Moonshine AI, a pioneering on-device voice AI company developing speech-to-text models that prioritize privacy, speed, and efficiency. Moonshine's models outperform OpenAI's Whisper while using significantly fewer parameters.

Pete was technical lead of the TensorFlow Micro team at Google and one of the original creators of the TensorFlow framework. He previously founded Jetpac, a deep learning startup acquired by Google in 2014, and worked as a senior engineer at Apple for five years.

Pete authored the leading textbook on embedded machine learning (TinyML) and created a standard dataset for short speech recognition. He holds a Bachelor's in Computer Science from University of Manchester and is pursuing a PhD at Stanford University.

Pete founded Useful Sensors in 2022, developing AI-enabled sensor modules for consumer electronics. Moonshine AI secured funding from Wing VC and IQT in July 2025 to accelerate development of on-device voice AI technology.

Moonshine's speech-to-text models achieve 6.65% word error rate with only 245 million parameters compared to Whisper Large v3's 1.5 billion parameters and 7.44% error rate. The company launched Moonshine Voice in February 2026, enabling streaming speech recognition with order-of-magnitude faster response times than alternatives.